π€ Computer Science
Generative UI & AI
2024λ
6μ 20μΌ
Next.js
GPT
reduxλ νλ¬Όκ° μνκ΄λ¦¬ λΌμ΄λΈλ¬λ¦¬μΈκ°?
semantic, syntactic, matrix factorization, window-based
Five different ways at creating the random seed on NLP
GPTμμ μλμμ± λΈλ‘κ·Έ λν λ΄μ
Performance dominance of deleting noised dataset on NLP
νΌλΈλ¦¬μ± λ¨κ³μμ λ‘λ©, μ€ν¨ μνλ₯Ό ν μ€νΈνκΈ° μν΄ νλ‘ νΈ mock μλ² λ§λ€κΈ°
This paper contributes on (1) simple human evaluation metric, (2) provide evidence correlation between perplexity and SSA, (3) end-to-end neural model with low perplexity has best performance.